Abstract: Falls prevention is an important evaluation index of hospital nursing quality, and also one of goals that are continuously concerned in the field of patient safety worldwide. Early identification of fall risks and nursing intervention can reduce the incidence of falls in hospitals. As a forecasting tool, the fall risk prediction model plays an important role in early prediction of fall risk, collecting data directly from electronic medical record system and modeling with machine learning as the main method is becoming a research hotspot in this field. This paper reviewed the research status of the fall risk prediction models at home and abroad from the perspectives of construction methods, data sources, model validation, aiming to provide reference for the construction of the fall risk prediction models for inpatients in China.